Multimodal-based machine learning approach to classify features of addiction using EEG
Poster No:
518
Submission Type:
Abstract Submission
Authors:
Ji-Yoon Lee1, Myeong Seop Song1, Woo-Young Ahn1, Jung-Seok Choi2
Institutions:
1Seoul National University, Seoul, Korea, Republic of, 2Samsung Medical Center, Seoul, Korea, Republic of
First Author:
Co-Author(s):
Introduction:
Addictions have recently been classified as substance use disorder (SUD) and behavioral addiction (BA), but the concept of BA is still debatable. Therefore, it is necessary to conduct further neuroscientific research to understand the mechanisms of BA to the same extent as SUD. The present study used machine learning (ML) algorithms to investigate the neuropsychological and neurophysiological aspects of addictions in individuals with internet gaming disorder (IGD) and alcohol use disorder (AUD).
Methods:
We developed three models for distinguishing individuals with IGD from those with AUD, individuals with IGD from healthy controls (HCs), and individuals with AUD from HCs using ML algorithms, including L1-norm support vector machine, random forest, and L1-norm logistic regression (LR). Three distinct feature sets were used for model training: a unimodal-electroencephalography (EEG) feature set combined with sensor- and source-level feature; a unimodal-neuropsychological feature (NF) set included sex, age, depression, anxiety, impulsivity, and general cognitive function, and a multimodal (EEG + NF) feature set.
·Fig. 1. Flowchart for calculating the accuracy of each algorithm and illustration of different feature set combinations based on cross-validation (CV) deviances
Results:
The LR model with the multimodal feature set used for the classification of IGD and AUD outperformed the other models (accuracy: 0.712). The important features selected by the model highlighted that the IGD group had differential delta and beta source connectivity between right intrahemispheric regions and distinct sensor-level EEG activities. Among the NFs, sex and age were the important features for good model performance.
·Fig. 2. Feature importance according to beta coefficients: comparison between internet gaming disorder (IGD) and alcohol use disorder (AUD)
Conclusions:
This is the first study to develop ML models to distinguish among patients with IGD, patients with AUD, and HCs using multimodal feature sets, including sensor- and source-level EEG and NFs. In particular, the changes delta and beta source-level FC within the right hemisphere can serve as a neurophysiological indicator for distinguishing between IGD and AUD. Notably, individuals with IGD and AUD have similar neuropsychological symptoms despite their dissociable neurophysiological mechanisms. Furthermore, the multimodal ML model for distinguishing IGD from HCs emphasizes the potential utility of ML models for diagnosing IGD. In conclusion, our findings enhance our understanding of the utility of ML techniques for detecting IGD based on neurophysiological and neuropsychological similarities and differences between IGD (a BA) and AUD (a SUD).
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Classification and Predictive Modeling
Connectivity (eg. functional, effective, structural)
EEG/MEG Modeling and Analysis 2
Multivariate Approaches
Keywords:
Addictions
Electroencephaolography (EEG)
Machine Learning
Modeling
Multivariate
1|2Indicates the priority used for review
Provide references using author date format
Choi, S. W. (2014). Similarities and differences among Internet gaming disorder, gambling disorder and alcohol use disorder: a focus on impulsivity and compulsivity. Journal of Behavioral Addictions, 3(4), 246-253. https://doi.org/10.1556/JBA.3.2014.4.6.
Kamarajan, C. (2020). Random Forest Classification of Alcohol Use Disorder Using EEG Source Functional Connectivity, Neuropsychological Functioning, and Impulsivity Measures. Behavioral Sciences, 10(3), 62. https://doi.org/10.3390/bs10030062.
Kinreich, S. (2021). Predicting risk for Alcohol Use Disorder using longitudinal data with multimodal biomarkers and family history: a machine learning study. Molecular Psychiatry, 26(4), 1133-1141. https://doi.org/10.1038/s41380-019-0534-x.
Lee, J. Y. (2022). Enhanced resting-state EEG source functional connectivity within the default mode and reward-salience networks in internet gaming disorder. Psychological Medicine, 52(11), 2189-2197. https://doi.org/10.1017/S0033291722000137.
Mumtaz, W. (2018). A review on EEG-based methods for screening and diagnosing alcohol use disorder. Cognitive Neurodynamics, 12(2), 141-156. https://doi.org/10.1007/s11571-017-9465-x.
Park, M. (2020). Neurophysiological and Cognitive Correlates of Error Processing Deficits in Internet Gaming Disorder. Cerebral Cortex, 30(9), 4914-4921. https://doi.org/10.1093/cercor/bhaa083.
Park, S. M. (2021). Identification of Major Psychiatric Disorders From Resting-State Electroencephalography Using a Machine Learning Approach. Frontiers in Psychiatry, 12. https://doi.org/10.3389/fpsyt.2021.707581.
Park, S. M. (2020). Respiratory sinus arrhythmia biofeedback alters heart rate variability and default mode network connectivity in major depressive disorder: A preliminary study. International Journal of Psychophysiology, 158, 225-237. https://doi.org/10.1016/j.ijpsycho.2020.10.008.
Park, S. M. (2017). Neural connectivity in Internet gaming disorder and alcohol use disorder: A resting-state EEG coherence study. Scientific Reports, 7(1), 1333. https://doi.org/10.1038/s41598-017-01419-7.
Son, K. L. (2015). Neurophysiological features of Internet gaming disorder and alcohol use disorder: a resting-state EEG study. Translational Psychiatry, 5(9), e628-e628. https://doi.org/10.1038/tp.2015.124.